import https://github.com/servo/html5ever

commit d1206daa740305f55a5fa159e43eb33afc359cb4

1 files changed, 81 insertions, 0 deletions

diff --git a/benches/html5ever.rs b/benches/html5ever.rs
new file mode 100644
index 0000000..ff20c4f
--- /dev/null
+++ b/benches/html5ever.rs
@@ -0,0 +1,81 @@
+#[macro_use]
+extern crate criterion;
+extern crate html5ever;
+
+use std::fs;
+use std::path::PathBuf;
+
+use criterion::{black_box, Criterion};
+
+use html5ever::tendril::*;
+use html5ever::tokenizer::{BufferQueue, Token, TokenSink, TokenSinkResult, Tokenizer};
+
+struct Sink;
+
+impl TokenSink for Sink {
+    type Handle = ();
+
+    fn process_token(&mut self, token: Token, _line_number: u64) -> TokenSinkResult<()> {
+        // Don't use the token, but make sure we don't get
+        // optimized out entirely.
+        black_box(token);
+        TokenSinkResult::Continue
+    }
+}
+
+fn run_bench(c: &mut Criterion, name: &str) {
+    let mut path = PathBuf::from(env!("CARGO_MANIFEST_DIR"));
+    path.push("data/bench/");
+    path.push(name);
+    let mut file = fs::File::open(&path).ok().expect("can't open file");
+
+    // Read the file and treat it as an infinitely repeating sequence of characters.
+    let mut file_input = ByteTendril::new();
+    file.read_to_tendril(&mut file_input)
+        .ok()
+        .expect("can't read file");
+    let file_input: StrTendril = file_input.try_reinterpret().unwrap();
+    let size = file_input.len();
+    let mut stream = file_input.chars().cycle();
+
+    // Break the input into chunks of 1024 chars (= a few kB).
+    // This simulates reading from the network.
+    let mut input = vec![];
+    let mut total = 0usize;
+    while total < size {
+        // The by_ref() call is important, otherwise we get wrong results!
+        // See rust-lang/rust#18045.
+        let sz = std::cmp::min(1024, size - total);
+        input.push(stream.by_ref().take(sz).collect::<String>().to_tendril());
+        total += sz;
+    }
+
+    let test_name = format!("html tokenizing {}", name);
+
+    c.bench_function(&test_name, move |b| {
+        b.iter(|| {
+            let mut tok = Tokenizer::new(Sink, Default::default());
+            let mut buffer = BufferQueue::new();
+            // We are doing clone inside the bench function, this is not ideal, but possibly
+            // necessary since our iterator consumes the underlying buffer.
+            for buf in input.clone().into_iter() {
+                buffer.push_back(buf);
+                let _ = tok.feed(&mut buffer);
+            }
+            let _ = tok.feed(&mut buffer);
+            tok.end();
+        })
+    });
+}
+
+fn html5ever_benchmark(c: &mut Criterion) {
+    run_bench(c, "lipsum.html");
+    run_bench(c, "lipsum-zh.html");
+    run_bench(c, "medium-fragment.html");
+    run_bench(c, "small-fragment.html");
+    run_bench(c, "tiny-fragment.html");
+    run_bench(c, "strong.html");
+}
+
+criterion_group!(benches, html5ever_benchmark);
+criterion_main!(benches);